A human color appearance model is so designed as to allow correctly predicting how a color is seen when a color chip having a field angle of 2° is given. FIG. 1 is a view showing the definition of a human visual field. A CIE 1931 standard colorimetric observer is applicable to the range of a field angle of 1° to 4°. Therefore, this applicable region is divided into a stimulus region having a field angle of 2° or less, an adjacent region having a field angle of 4° or less, a background region from the adjacent region to a region having a field angle of 10°, and a surrounding region around this background region. Also, a visual field region including all these regions is an adaptation region.
CIE CAM97s is a representative color appearance model, and in this model the following can be set as viewing condition parameters.
La: Absolute luminance [cd/m2] in adaptation region                Normally, 20% of white point absolute luminance in adaptation region        
XYZ: Relative XYZ value of color chip
XwYwZw: Relative XYZ value of white point
Yb: Relative luminance of background region
Surround conditions:                Average Surround (larger than a field angle of 4° of a color chip)        Average Surround (equal to or smaller than a field angle of 4° of a color chip)        Dim Surround        Dark Surround        Cut-Sheet Transparencies (on viewing box)        
The surround condition is Average if the relative luminance in the surrounding region is 20% or less of a white point in the adaptation region, Dim if this value is smaller than 20%, and Dark if this value is almost 0%.
A color appearance model is derived from experimental results using monochromatic color chips. Hence, no method has been established which determines viewing condition parameters applicable to an image having a plurality of colors. That is, the relative luminance Yb of the background region is set at 20% because neutral gray is 20% of a white point.
Also, when a color appearance model is applied to an image, one viewing condition parameter is generally used for all pixels.
When one viewing condition parameter is thus applied to all pixels, a visual effect between a single color and a background expressed on a rasterized image cannot be reflected on a color matching result.
Furthermore, since an average viewing condition parameter is evenly applied to an image, no color matching result having high accuracy can be locally obtained.